HAL Id: tel-00493130
https://tel.archives-ouvertes.fr/tel-00493130
Submitted on 18 Jun 2010
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
REACTIONS
Sonia Pin
To cite this version:
Sonia Pin. PROBING THE INITIAL STAGES OF SOLID STATE REACTIONS. Chemical Sciences.
Université Joseph-Fourier - Grenoble I; Università degli studi di Pavia, 2010. English. �tel-00493130�
Alma Ticinensis Universitas – Università degli Studi di Pavia Université Joseph Fourier de Grenoble
Sonia PIN
DOTTORATO DI RICERCA in Scienze Chimiche XXII CICLO
Settore scientifico disciplinare: CHIM02
ECOLE DOCTORALE Matériaux et génie des procédés ECOLE DOCTORALE N°228
Relatore (Italy): Prof. Paolo GHIGNA – Università degli Studi di Pavia
Directeur de thèse (France) : Prof. Michel DUCLOT – Université Joseph Fourier
Studio dei primi istanti delle reazioni solido - solido
Probing the initial stages of solid state reactions Etude des instants initiaux des réactions
solide – solide
!
" " #" $ "
Riassunto
Il presente lavoro di tesi di dottorato intitolato “Studio dei primi istanti delle reazioni solido – solido” è stato realizzato nel quadro di una tesi in co-tutela franco- italiana sotto la doppia direzione dell’Alma Ticinensis Universitas – Università degli Studi di Pavia e dell’Università Joseph Fourier di Grenoble. Si è svolto principalmente al Dipartimento di chimica – fisica “M. Rolla” dell’Università di Pavia, sotto la supervisione del Prof. Paolo GHIGNA. La parte francese si è svolta sotto la supervisione del prof. Michel DUCLOT al Laboratoire de Electrochimie et de Physico- Chimie des Matériaux et des Interfaces – LEPMI (UMR 5631, CNRS, Grenoble-INP, UJF). Inoltre, i lavori sperimentali di studio della cinetica delle reazioni tra ossidi sono stati svolti presso lo European Synchrotron Radiation Facility (E.S.R.F., Grenoble, France), sotto la supervisione del Dr. Francesco D’ACAPITO.
La tesi è organizzata in sette capitoli. L’introduzione fornisce una visione generale sulle conoscenze delle reazioni allo stato solido, sia dal punto di vista termodinamico che da quello del trasporto della materia. Il primo capitolo fornisce il punto della conoscenza sulle reazioni solido-solido e mostra l’interesse per la conoscenza degli istanti iniziali. Proprio questo punto costituisce l’originalità del lavoro sviluppato. Nel capitolo 2 sono presentate le strategie di analisi dei dati e il tipo di modello sperimentale (film sottile di ossido depositato su un substrato monocristallino).
Il capitolo 3 è dedicato alla crescita epitassiale del film sottile di ossido tipo MO, (M = Ni, Zn, Mn), depositato mediante RF-magnetron sputtering. Il campione così preparato è poi cotto in modo da avere una buona orientazione del film.
La tecnica di X-ray Absorption Spectroscopy (XAS) e Fluorescence X-ray Absorption Spectroscopy (Fluo-XAS) sono presentati nel capitolo 4 come la beamline GILDA BM08 di ESRF, dove sono stati svolti gli esperimenti che hanno permesso l’identificazione di almeno un nuovo composto cristallino. Sulla beamline ID24 (ESRF) sono state realizzate le misure di micro X-ray Absorption Near Edge Spectroscopy (micro-XANES) (capitolo 5). Sulla beamline ID03 (ESRF) sono invece state svolte le misure di diffrazione superficiale (capitolo 6).
Un passo avanti nell’analisi della superficie viene effettuato nel capitolo 7 dove le misure di Atomic Force Microscopy (AFM) e di Diffrazione Elettronica (ED) mostrano impressionanti cambiamenti della morfologia superficiale dei film di ossido (capitolo 8).
Questo lavoro ha permesso un approccio innovativo per la comprensione dei meccanismi delle reazioni interfacciali a livello atomico. Questi risultati possono essere un contributo importante per l’elaborazione di nuovi ossidi a livello nanometrico.
Questa prospettiva lascia pensare all’apertura di una nuova via per lo studio dei meccanismi di reazioni allo stato solido in modo analogo alla femtochimica in fase gassosa.
Abstract
The work presented in this PhD thesis and entitled “Probing the initial stages of solid state reactions” has been mainly carried out under a Convention de Co-tutelle de thèse between France and Italy and the two universities, Alma Ticinensis Universitas – Università degli Studi di Pavia and Université Joseph Fourier de Grenoble. The work has been mainly carried out at the Department of Phisical – Chemistry “M. Rolla” of the University of Pavia under the supervision of prof. Paolo GHIGNA and the Laboratoire de Electrochimie et de Physico-Chimie des Materiaux et des Interfaces – LEPMI (UMR 5631, CNRS, Grenoble-INP UJF) under the direction of Prof. Michel DUCLOT. The work of study of the kinetic of the reaction between oxides was performed at the European Synchrotron Radiation Facility (E.S.R.F., Grenoble, France), under the supervision of Dr. Francesco D’ACAPITO.
The thesis is organized in seven chapters. The introduction presents the general concept of solid state reactions, under the point of view of the thermodynamic and of the matter transport. The first chapter shows what we know about reactions in the solid state and the high need for the understanding of the early stages nowadays. It’s this crucial point the one that shows the peculiarity of this work. In chapter 2 the analysis strategy and the model that have been choose (thin film onto a monocrystal) are presented along with the main options exploited during the PhD.
Chapter 3 is dedicated at the epitaxial growth of the thin film of oxide MO, (M = Mn, Ni, Zn), deposited by means of RF-magnetron sputter and subsequently the anneal and the reactions are carried out to have a good film orientation.
The X-ray Absorption Spectroscopy (XAS) technique and the Fluorescence X- ray Absorption Spectroscopy (Fluo-XAS) with the GILDA BM08 beamline (ESRF) are presented in chapter 4, whit the identification of at least a new crystalline compound.
The ID24 beamline (ESRF) has been used for the micro X-ray Absorption Near Edge Spectroscopy (micro-XANES) experiment (chapter 5). The ID03 beamline (ESRF) has been used for the surface diffraction measurements (chapter 6).
A step forward in the surface analysis is presented in chapter 7 where the Atomic Force Microscopy (AFM) and Electron Diffraction (ED) techniques reveal an impressive change in the surface morphology of the reacted films (chapter 8).
This work has involved a completely new approach for the comprehension of the mechanisms in the interfacial reactions at atomistic level. These results may be an important contribution for the design of new oxides at nanometric scale. This perspective open a new way for the study of the mechanisms in solid state reactions similarly to femtochemistry in gas phase.
Résumé
Le présent travail intitulé « Etude des instants initiaux des réactions solide – solide » a été réalisé dans le cadre d’une thèse en co-tutelle franco-italienne sous le double sceau de Alma Ticinensis Universitas de Pavie et de l’Université Joseph Fourier de Grenoble. Il s’est déroulé pour l’essentiel au ‘Département de Chimie-physique Mario Rolla’ de l’université de Pavie, sous la direction du Prof. Paolo GHIGNA. La partie française c’est déroulé sous la direction du Prof. Michel DUCLOT au Laboratoire d'Electrochimie et de Physico-Chimie des Materiaux et des Interfaces - LEPMI (UMR 5631, CNRS, Grenoble-INP, UJF). De plus, les travaux expérimentaux du suivi de la cinétique des réactions oxyde - oxyde ont été réalisés au ‘European Synchrotron Radiation Facility’, (E.S.R.F, Grenoble, France), sous la direction du Dr. Francesco D'ACAPITO.
Le mémoire est subdivisée en 7 chapîtres. L’ introduction fait un point général sur les connaissances des réactions entre solides tant du point de vue thermodynamique que du point de vue transport de la matière. Le première chapitre fait le point des connaissances de la cinétique des réactions solide – solide et montre l’intérêt de la connaissance des phénomènes aux instants initiaux. Ce point constitue la partie originale qui sera développé dans ce travail. Dans le chapître 2 sont présentés la stratégie d'analyse (rayonnement synchrotron, ajustement des données) ainsi que le système expérimental (film mince du matériaux étudie sur substrat monocristallin).
Le chapître 3 est consacrée a l’étude de la croissance épitaxiale des films minces d’oxyde de type MO, (M = Mn, Zn, Ni), à l’aide d’un système de déposition par
magnetron sputtering Radio Frequence (RF-magnetron sputter). L’ échantillon obtenu est ensuite recuit pour l’homogéneité de l’orientation épitaxiale.
Les techniques X-ray Absorption Spectroscopy (XAS) et Fluorescence X-Ray Absorption Spectroscopy (fluo-XAS) sont présentées dans le chapître 4 ainsi que le faisceau synchrotron GILDA BM08 de l’ESRF, où ont été réalisées les expériences ce qui a permis l'identification d'au moins un nouveau composé cristallin. Sur le faisceau ID24 (ESRF) ont été réalisées les mesures de X-ray Absorption Near Edge Spectroscopy (micro-XANES) (chapître 5). Le faisceau ID03 (ESRF) a servi pour les mesures de diffraction superficielle (chapître 6).
Un pas en avant dans l'analyse de la surface est présenté dans le chapître 7 où les mesures par Atomic Force Microscopy (AFM) et par Electron Diffraction (ED) montrent des changements impressionnants de la structure et de la morphologie superficielle des films d’oxyde (chapître 8).
Ce travail a permis un approche performante pour la compréhnsion des mécanismes des réactions interfaciales au niveau de l’atome. Ces résultats peuvent être une contribution important pour l’élaboration de nouveaux oxydes à l’état nanometric.
Cette perspective laisse penser à l’ouverture d’une nouvelle voie pour l’étude de mécanismes de réaction à l’état solide de façon analogue au réactions de femtochimie en phase gazeuse.
I NDICE
Introduzione……… 1
1 – Reazioni Solido-Solido: la necessità di conoscere “Il principio”….…… 9
1.1 – Reazioni solido-solido: introduzione….………. 9
1.2 – Interfacce………. 11
1.3 – Reazioni topochimiche………. 12
1.4 – Perchè i primi istanti?... 14
1-5 – Riferimenti...………... 17
2 – La ‘Caccia’ alla reazione....…..……….…...… 19
2.1 – Primo step: la reazione.…….….……….……… 19
2.2 – Secondo step: lo spinello....…….……… 20
2.3 – Il design sperimentale………..……….……... 24
2.4 – Riferimenti...………...…. 26
3 – Un modello per esplorare esperimenti e teoria………….………. 28
3.1 – Costruendo i primi istanti del modello...……….…….. 28
3.2 – Deposizione del film sottile di ZnO……….………..…… 29
3.3 – Riferimenti..……… 37
4 – G.I.L.D.A. – BM08 a E.S.R.F. e la fluorescence-XAS……….. 38
4.1 – XAS e Fluo-XAS: chi sono?... 38
4.2 – L’equazione dell’EXAFS..……….. 41
4.3 – La tecnica sperimentale...……….. 45
4.4 – Strategia di fit..……… 46
4.5 – Parte sperimentale……… 48
4.6 – Riferimenti……….. 59
5 – ID24 a E.S.R.F. e le Micro-XANES…...……….. 60
5.1 – Cosa sappiamo delle XANES?... 60
5.2 – Parte sperimentale..……….………. 64
5.3 – Riferimenti...……….………... 71
6 – ID03 a E.S.R.F. e la diffrazione superficiale aiutano la tradizionale XRD………. 72 6.1 – Diffrazione superficiale: la sorella minore della tradizionele XRD……… 72
6.2 – Parte sperimentale.……….. 75
6.3 – Riferimenti……….. 79
7 – AFM ed ED: il modo per vedere l’invisibile……….………. 80
7.1 – Atomic Force Microscopy parte sperimentale..……….. 80
7.2 – Misure ED parte sperimentale…...……… 84
8 – Prima ipotesi di meccanismo per l’interfaccia reattiva
(
0001)
ZnO ||( )
1120 sapphire... 86Conclusioni…….……….… 93
Appendce A……….. 97
I NDEX
Introduction……… 1
1 – Solid state reactions: the needs to know “the begin”…………..…………. 9
1.1 – Solid state reactions: introduction…..……… 9
1.2 – Interfaces………..….. 11
1.3 – Topochemical reactions….……… 12
1.4 – Why the first stages?... 14
1-5 – References……….…. 17
2 – The “hunt” for the reaction………..……….. 19
2.1 – Step one: the reaction……….……… 19
2.2 – Step two: the oxide spinel………...……… 20
2.3 – The experimental design…...……….. 24
2.4 – References……….….. 26
3 – A Model to explore experiment and theory…………...……… 28
3.1 – Building up the first stages of the model…………...………. 28
3.2 – ZnO thin films deposition…………...………..…….. 29
3.3 – References………..……… 37
4 – G.I.L.D.A. – BM08 at E.S.R.F. and the fluorescence-XAS…….…………. 38
4.1 – XAS and Fluo-XAS: who are them?... 38
4.2 – The EXAFS equation……….……… 41
4.3 – The experiment technique……..………. 45
4.4 – Fit strategy……….……….. 46
4.5 – Experimental part……… 48
4-6 – References………..…. 59
5 – ID24 at E.S.R.F. and the micro-XANES……… 60
5.1 – What do we know about XANES?... 60
5.2 – Experimental part……… 64
5.3 – References………..…. 71
6 – ID03 at E.S.R.F. and the surface diffraction help the traditional XRD…. 72 6.1 – Surface diffraction: the young sister of the traditional XRD………. 72
6.2 – Experimental part……… 75
6.3 – References...………...…. 79
7 – AFM and ED: the way to see the invisible………..………... 80
7.1 – Atomic Force Microscopy experimental Part…….……… 80
7.2 – ED measurements experimental part……….. 84
8 - Topotactical relations in
(
0001)
ZnO ||( )
1120 sapphire reactive interface: a first hypotesis of mechanism………. 86Conclusions………. 93
Appendix A …..………...… 97
Sommaire
Introduction……… 1
1 – Réactions solide-solide: la nécessité de connaître "les premier instants" 9 1.1 – Réactions solide-solide: introduction……… 9
1.2 – Interfaces……….…… 11
1.3 – Réactions topochimiques.…..……… 12
1.4 – Pourquoi les premiers instants ?... 14
1-5 – Bibliographie...………... 17
2 – La ‘chasse’ à la réaction..………... 19
2.1 – Première étape : la réaction.………. 19
2.2 – Seconde étape: le spinel…....……… 20
2.3 – Le design expérimental.……….…. 24
2.4 – Bibliographie..……… 26
3 – Un modèle pour explorer essais et théorie………. 28
3.1 – En construisant les premiers instants du modèle………..…..……….. 28
3.2 – Déposition du film mince de ZnO……….……..…….. 29
3.3 – Bibliographie..………..….. 37
4 – Faisceau G.I.L.D.A - BM08 à E.S.R.F et le Fluorescence-XAS…………... 38
4.1 – XAS et Fluo-XAS: qu’est ce ?... 38
4.2 – L'équation de l'EXAFS..………. 41
4.3 – La technique expérimentale….………. 45
4.4 – Stratégie de l’ajustement.……….. 46
4.5 – Partie expérimentale..……….……… 48
4-6 – Bibliographie.……….…… 59
5 – Faisceau ID24 à E.S.R.F et elle Micro-XANES.……… 60
5.1 – Que savons nous savons des XANES ?... 60
5.2 – Partie expérimentale.……….……… 64
5.3 – Bibliographie.……….……... 71
6 – Faisceau ID03 à E.S.R.F et la diffraction superficielle aident les
traditionnels XRD………. 72 6.1 – Diffraction superficielle: la ‘petite sœur’ de traditionelles XRD………….. 72
6.2 – Partie expérimentale………... 75
6.3 – Bibliographie………..……….…... 79
7 – AFM et ED: ‘la manière’ pour voir l'invisible.…..……… 80 7.1 – Atomic Force Microscopy (AFM) : partie expérimentale……..………….. 80 7.2 – Mesures par Electron Diffraction (ED) : partie expérimentale....………….. 84 8 - Première hypothèse de mécanisme pour l'interface réactive
(
0001)
ZnO ||( )
1120 sapphire ……….……… 86Conclusion……….…….. 93
Annexe A……….……….…... 97
I NTRODUCTION
“The scientist only imposes two things, namely truth and sincerity, imposes them upon himself and upon other scientists”
(E. Shrodinger)
The assessed theory of the kinetics of solid state reactions essentially describes the growth of a product phase enclosed between two grains of the reagent phases. In this treatment, various kinetic processes are considered, depending on the nature of reagents and product, the process variables, and the extent of the reaction [1]. In particular, as the reaction proceeds and the product layer becomes thicker, local chemical equilibrium is possibly attained at each interface so that the growth becomes kinetically controlled by diffusion in the product layer. Then, the theory predicts a parabolic growth law [2]
(squared thickness proportional to time) and relates the experimentally accessible rate constant (the proportionality factor) to quantities that can be independently determined.
Different mechanisms can be assumed depending on the structure elements or defects that are actually diffusing in the product phase, and the theory allows identification of the mechanism by comparing the rate constant to the diffusion coefficients of the mobile species or conversely to predict a reaction rate from the thermodynamic and transport properties of the product phase.
Useful additions to this general treatment take into account kinetic control by other processes, such as heat transport or interface mobility. The latter case is
particularly interesting as it extends the theoretical approach towards the initial stages of the reaction, when the product layer is thinner and the growth rate is constant (linear growth).
In both cases, i.e. diffusion control or interface mobility control, a basic assumption is that a product layer already exists. As matter of fact, practically nothing is presently known about the very early stages of the process, when the product layer is not yet formed and the chemical reactivity is entirely controlled by interfacial aspects.
This quite surprising lack of basic understanding is related to the fact that the subject is largely unexplored mainly because of the lack of an established probe.
Solid state reactivity in its various aspects plays indeed a significant role in many different areas of science, such as Crystallography, Mineral Physics and Earth Sciences, Physical Metallurgy, Ceramic Science, Semiconductor Physics, Solid State Chemistry, but the basic knowledge and the main results (not to speak of the language itself) are not always well shared among the disciplines. Surface Science and Catalysis provide an ample background of knowledge on the transformations and reactions occurring at a free interface, while the processes occurring at the boundary between two condensed phases are an important topic in many of the scientific areas above quoted, for instance in the investigation of phase transformations or growth of thin films. All this knowledge provides important starting points but does not directly face the processes occurring when a single interface between two grains of the reagent phases is progressively turned into two different interfaces between each reagent and the newly
reactions involving two condensed phases as reagents and producing a different crystalline phase, the present state of art is that thermodynamics gives tools to predict whether or not the formation of a particular product can occur, and the assessed kinetic theory gives tools to predict the growth rate of this product when diffusion is the rate determining step. However, we do not really understand the mechanism through which this product is formed. This means that we have no general knowledge of the very basic aspects of the chemical reactivity in the solid state; for instance we ignore how and why a reaction actually goes towards one or another polymorph, a compound or a broad range solid solution, a stable or metastable product.
Quite impressively, the scientific community still needs an assessed agreement concerning the procedures and the aforementioned experimental probes required to investigate this topic, as well as a sound common background connecting the different areas that can seemingly provide important contributions to the advancement of knowledge, and that conversely can profit from that advancement.
Aim of the thesis is therefore to reach a basic understanding of the very initial stages of solid state reactions by organizing into a work team scientists with different backgrounds, in particular by adding to the core expertise in crystallography the required knowledge concerning solid state reactivity and defect chemistry, as well as the most effective characterization techniques, including those based on synchrotron radiation. We want to make clear that the main objective is to contribute to the advancement of knowledge in a very basic but somehow unexplored area of science which is at the crossroad of many disciplines. The objective can profit from many
sparse insights on the role of interfacial reactive steps so far available in the scientific literature on topics that are differently related to the subject of the present work, and is eventually achieved by combining into an original methodology the theoretical approaches and experimental tools coming from those different scientific areas.
The thesis is expected to have a strong impact on Mineral Physics and Earth Sciences. Geological history has been deciphered so far mainly by looking for markers of P-T-t paths in rocks, i.e. physicochemical responses of the minerals to, and fingerprints in minerals of, P-T-t changes. Microstructures and microtextures, which concern the small-scale arrangements of mineral species in rocks, are interesting in this respect because can be used in tracking back the conditions of such changes in view of the transformation mechanisms they suggest. It is their non-equilibrium nature that makes them particularly useful in tracking mineral behaviour.
Topological studies of microstructures and nanostructures (nature, size, mutual orientation of crystals, shape and orientation of their interface, etc.) of frozen-in mineral reactions provide significant evidence and are carried out to determine the types of reactions taking place in rocks. However, the interpretation of experimental data is subject to uncertainties. The recognition of the clues and evidences is tributary to the strength of the link that exists between the object and the mechanism. The present thesis is expected indeed to strengthen such link by giving new and fundamental insight into the mechanisms of solid state reactions.
The thesis is also expected to have a strong impact on the methods of preparing new crystalline materials or, more generally, on solid state synthesis. Generally speaking, this seems quite obvious, but actually requires a few comments as many areas of the current research on new materials do not clearly show such a strong demand for a better understanding of the very basic aspects of solid state reactivity. In this area, the synthetic activity actually receives far less attention than property characterization, and is frequently done on a very empirical basis; a rational approach, when present, is mostly oriented to understanding the thermodynamic aspects or optimizing a reaction rate. The nature of the reaction product appears quite obvious in almost all cases, so that there is seemingly no interest in understanding how and why a given product nucleates.
When the products have more complex crystal structures and composition, the chemical reaction requires a significant reorganization of the structure elements, and the situation changes. Although not strictly pertinent to the topic under discussion, a very impressive example of the role of structural complexity in controlling the nature of a product is the production of a metallic glass from fast cooling of a melt. This process is essentially based on hindering crystallization of thermodynamically more stable compound phases, which is found to become progressively more difficult as more components and complex intermetallic structures are involved.
To sum up, the interest for a better understanding of the very early stages of solid state reactions is a current issue in the mineralogical community and is expected to propagate rapidly to materials science and solid state synthesis, where the interest is seemingly confined at present to specialized areas. Concerning the high temperature oxide superconductors, for instance, an ample variety of increasingly complex structures are under investigation, and current research already faces the difficulty of planning an
effective synthetic protocol using conventional solid state synthesis (sometimes known as “ceramic” synthesis). The same situation is expected in the near future in many other areas of materials science and solid state synthesis, as more complex crystalline phases with many different structural sites and complex stoichiometries (even neglecting doping) are progressively receiving attention (multiferroic materials and materials for spintronics being among the most pertinent examples) because of the very sophisticated functional properties so achieved or the possibility of combining in a single phase functional properties peculiar of significantly different compounds.
We finally note that understanding how the interfaces between condensed phases and their time evolution affect chemical reactivity is nanoscience by itself, so that it is almost impossible to quote all the expected connections of this research thesis with nanosciences and nanotechnology. In particular, a better understanding of these aspects is expected to provide strong impact on the knowledge of the chemical mechanisms underlying the deposition of thin films on various substrates, a preparative technique that spans a broad range of applications such as catalysis, photoluminescence, spintronics, information storage (magneto-resistance and related properties).
To introduce the central topic of the thesis, we note that the crucial role played by the interfacial free energy in heterogeneous reactions has been evidenced mainly by means of local inspections (for example, by means of TEM) of interfaces between bulky materials. It is well established that nucleation and growth of the product phase from a couple of reactants in the solid state is constrained by more factors than ‘free’
FIG. 1 Effect of the interfacial free energy ( ) on the relative stability of stable and metastable compounds (referred to the bulk, lower part of the diagram). The dashed lines refer to the less stable phase both in the bulk and in presence of interfacial free energy. The reactive system in this regime is also depicted.
phases controls the interfaces which in turn control the growth of the new phase and its spatial and orientational relationships with the parent phases.
In the case of a tight seed/host crystal boundary, the lowest activation energies of nucleation are realized (i) for a specific orientation of the seed to the host (topotaxy) and (ii) in the case of minimum stress/strain along and across the interface boundaries.
A good coincidence of translational lattices and similarities in their atomic structures are in fact expected to favour nucleation since the reactant/product interface energy is lowered.
Because of the critical role of interface, it is possible to conceive to maximise and tune its effect. First, the role of the interface can be maximised by using one of the reactant in form of films of different thicknesses. Second, three different regimens can be explored and clarified:
1. For the very thin films, the reactive system is quasi-2D: the interfacial term dominates the overall free energy. Different products can be thermodynamically stabilised by the interfacial free energy: for given structures and compositions the effect is controlled by the topotactical relationships and is different for different interfaces (Fig. 1a). This has been recently evidenced by a preliminary study by the proposing team.
FIG. 2 In the diffusion controlled regime, the reactive system is shown with reactants and product layer in between.
2. For intermediate thickness, once the final phase has nucleated, interfacial movement is the rate determining step: topotaxy and interfacial free energy control the rate of the reaction.
3. For thick reaction layers, diffusion is the rate determining step and the interfacial term is negligible (Fig. 1b).
In other terms, with respect to the bulk contribution, the interfacial contribution to the free energy of a layer decreases as the thickness increases. We can switch from a situation where the interfacial contribution is the largest to another where bulk contribution rules the process, that is from a quasi–2D to a 3D reactive system. We expect that the transitions between the regimes could be directed towards different products and investigated by changing the dimensionality of the reactive couple. To the best of our knowledge, this approach towards the reactivity in the solid state is completely original.
R EFERENCES
[1] M. Cournil and G. Thomas (1982) J. Chim. Phys. 79, 729.
[2] M. Cournil and G. Thomas (1977) J. Chim. Phys. 74, 545.
1. S OLID S TATE R EACTIONS : T HE N EED T O K NOW “T HE B EGIN ”
“There are two ways to live your life.
One is as though that nothing is a miracle.
The other is as though everything is a miracle.”
(A. Einstein)
1.1 - S OLID S TATE R EACTIONS: I NTRODUCTION
The concept of Solid State Reactions interests and inspires scientists worldwide.
Solid State Reactions (SSR) [1-3] are of extreme interest both for technological applications and from the point of view of basic science. On one side, many important materials are currently prepared by solid state synthesis [4], for example almost all ceramic materials; on the other side, reactions in the solid state occur in many geological processes such as mineral metamorphism.
Nowadays the idea of reactions in the solid state is widespread in chemistry and related disciplines. We can refer to solid state reactions with “modern” words and concepts. A solid state chemical reaction in the classical sense occurs when local transport of matter is observed in crystalline phases. This definition does not means that gaseous or liquid phases may not take part in solid state reactions. However, it does means the reaction product occurs in a solid phase [5].
If the reactants are brought together at a constant pressure and temperature in a closed system, than the reaction will takes place spontaneously if the Gibbs free energy
of the system is decreased. During the course of a solid state reaction, there are local changes of the chemical potentials of the individual components i of the crystal. In the narrow sense, solid state reactions are characterized by the diffusion of atoms of type i (or of ions of type i) with charge zi in chemical or electrochemical potential gradient.
Thus, the local changes of partial molar free energies of the various type of particles are the driving forces for chemical reactions in the solid state and the rate of diffusion is proportional to the driving force by means of a factors of proportionality called transport coefficient.
The transport of matter in the solid state, and thus also the reactivity of solids, are dependent upon the mobility of the individual particles in the structure. Every case of mass transport in solid phases is directly dependent upon deviation from ideal crystalline order. The rate constants in crystalline phases are directly related to the atomic disorder and the higher atomic disorder, the higher is the corresponding transport coefficient.
Solid state reactions are divided into the following groups:
1. homogeneous reactions
2. reactions in single phase inhomogeneous systems 3. heterogeneous reactions
Chemical reactions between solid crystalline materials are in general exothermic because of the high degree of order of crystalline phases, which means that the entropy difference between these phases is relatively small.
If two substances react with one another to form one or more product phases which are separated from the reactants and from other by phase boundaries, then a heterogeneous solid state reaction is said occur [6]. The simplest product phases are binary, as in the reaction equation
nA + mB AnBm
The reaction product separate the reactants from one another, and the reaction proceeds by diffusion of the participating components through the reaction product. For very low solubility of the reactants in the reaction product the particle fluxes are locally constant, and as long as local thermodynamic equilibrium is maintained at the phase boundaries, a parabolic growth law results.
From a single crystal of AO and a single crystal of B2O3, a partially monocrystalline spinel AB2O4 is formed and in this case the fluxes of the components in the reaction product, responsible for the advancement of the reaction, are fluxes of charged particles. Therefore, in order to preserve local electrical neutrality, the fluxes of different ions must always be coupled with each other. Consequently, the following combination are possible: (i) either oppositely charged ions flow in the same direction;
(ii) ions with like charges flow in opposite directions through the reaction product.
These simple considerations give the possible limiting cases of a solid state reaction between ionic crystals.
1.2. - I NTERFACES
Interfaces play an important role in solid state reactions and, as site of repeatable growth, they can permit equilibrium between point defects to be attained. At
thermodynamic equilibrium, the electrochemical potential across an interface remains constant, but the chemical potentials of the components change because of change in structure. Therefore, electrical charges and discontinuities in electrical potential are observed. A measure of penetration depth of the electrical space charge layer is the so- called Debye-Hückel length:
2 / 1
2
8 0
= ne
l kT π i
ε (1-1)
Where is the relative dielectric constant, and ni is the number per cm3 of univalent particles with positive or negative excess charges in the lattice. If several different sort of mobile particles i are present, ni in the denominator must be replaced by
i2 iz
n . The Debye-Hückel length here is taken from the theory of electrolyte solutions. A large number of interfacial phenomena are associated with the electrical surface of interface charge layer.
In order to bring atoms reversibly and isothermally from the undistorted interior of a crystal to an interface an expenditure of free energy is required.
1.3. - T OPOCHEMICAL R EACTIONS
Topochemical reactions are solid state reactions under action of a chemical potential gradient that occur essentially at distinguished sites reactant. There are a large number of classical solid state reaction as well as tarnishing reactions in which the morphology of the reaction product is result of the existence of fast transport paths in
which the transport coefficient is relatively high. The morphologies can often be quite unusual and interesting.
Let us consider the formation of spinel in the reaction couple AO/AB2O4/B2O3. It was assumed that the mobility of the oxygen ions is negligible. However, if the reactant oxides AO and B2O3 are very porous, then the reaction product AB2O4 will also be porous. Furthermore, if the reaction product is an electronic conductor, as is generally the case when the ions A and/or B are transition metal cations, then the oxygen can be transported via the gas phase through the pores, while diffusion of electrons (or electron holes) and ions occur simultaneously in the reaction product.
The reaction rate will then be determined by the diffusion of the faster of the cations, since the transport number of the electronic charge carriers is essentially unity, and, for a sufficiently high porosity, the oxygen gas will always be available everywhere at the reaction front.
Even for compact product layers, gas transport can still occur at a free surface, and this can lead to a unique morphology [7].
Here again, the electronegative component is transported rapidly via the gas phase, while the electrons are transported through the solid, provided that the solid is an electronic conductor. The reaction product can then grow parallel to the surface. The rate of this growth is controlled by the diffusion of the faster cation, since the slower cation is always available for reaction at the three-phase boundary gas/reaction product/reactant. Therefore, as a rule, the overlap of the reaction product at the surface will always occur in the direction towards the reactant containing the slower cation (B
in this case). From the ratio of product layer thicknesses ∆x1 ∆x2 , the ratio of the diffusion coefficients of the slower and faster cations can be calculated. If
+ +
− << 3 << 2
2 B A
O D D
D in the reaction product, then it follows:
2 / 1 2
/ 1
2 1 2
1 ≈ = 8
∆
∆
A B
D D k
k x
x (1-2)
The numerical factor 8 in Eq. (1-2) reflects the fact that the reaction mechanism at the surface and in the interior are different, do that reaction volumes for the passage of one equivalent are not the same in each case.
Surface diffusion can also provide a fast transport path, and than can thus also give rise to rapid growth along the surface.
1.4. - W HY THE F IRST S TAGES?
The kinetics and mechanisms of solid state reactions are typically studied making use of mono-dimensional chemical diffusion experiments.
A parabolic rate law is experimentally observed, and this can easily be explained as follows. First of all, we may assume that the local equilibrium occurs at the phase boundaries. This means that, at the phase boundaries, all the thermodynamic variables are fixed, and so the local defect concentrations are also fixed for all time. Therefore, an average concentration gradient of defects is formed in the reaction layer which is
determining ionic flux is therefore given by ji ∝ 1/ x. Since ji is proportional to the instantaneous growth rate x/dt of the layer, it follows that d x/dt ∝ 1/ x. This expression may be integrated to give the parabolic growth law [8, 9] in the form:
x2 = 2k t (1-3)
Where k is known as the practical reaction rate constant.
However, the mechanisms and the kinetics of the interfacial reactions at the first stages (that are, in a sense, the ‘true’ chemical reactions) are completely unknown, when local chemical equilibrium is not yet obtained [10] and the chemical kinetics is not driven by long range diffusion.
The initial state at an interface between two solid substances with tendency for compound formation is a strong non-equilibrium situation. This is in contrast to later stages of solid-state reactions where compound formation has already be formed, resulting in reduction of the driving forces for subsequent reactions. F. M. d’Heurle [11]
and R. Bormann [12] reported that from calculations based on the large driving force available in many systems for first-phase formation is shown that the barrier for nucleation of the first phase is very small, typically the size of the critical nucleus is less than one unit cell. The scientists agree that this very small critical nucleus size indicates that either the application of classical nucleation theory is invalid in this case or that the nucleation site density is so large that the volume fraction associated with nucleation and growth to impingement of the first layer of grains of the product phase is immeasurably small. They suppose that in either case, nucleation should be excluded as a rate-determining factor in the initial reaction between elements and one-dimensional
growth of the product phase in the direction perpendicular to the interface should occur from the very beginning of the reaction.[13-15]
This observation, however, cannot explain some experimental observations. For example, numerous studies of binary systems demonstrated that certain phases are favoured as the first phase, even though the free energies of formation of all the compounds are large enough that anyone could be the first phase. Second, the first phase to form is frequently a metastable phase. The formation of metastable phases requires the presence of barriers, such as interfacial reaction barriers [16, 17] to the formation of the competing equilibrium phases.
From this angle, solid state reaction kinetics (and mechanisms), at the early stages, is (are) the least developed field of chemical kinetics (and mechanisms), in which the gap between theory and experiment seems to be greatest. Introducing the concept of a simple model reaction seems to be a crucial point for further developing the theory towards deeper insight into chemical mechanisms and kinetics at the first stages of solid state reactions, in fact there are currently just few examples of reaction models for solid state reactions at these stages.
The gap between experiments and theory in the early stages of solid state reactions is principally due to the need of an adequate probe to follow these processes (when the equilibrium at the two interfaces is not yet attained as said at the begin of this chapter and the traditional techniques are not applicable [18, 19]) and of a standardized procedure to follow.
An intensive development of Synchrotron Radiation (SR) sources, during the last twenty years, made routine its various applications in chemistry and makes SR techniques really sophisticated and useful tool in the investigation of the early stages of solid state reactions and for the comprehension of all these experimental evidents that cannot be explained with the classical theory , as we’ll see more in detail in the next chapters.
It’s clear that, once we will be able to fully understand solid state reactions from the early stages, we will be able to control them from the beginning and this important point will make easier the synthesis at the nanoscale and the comprehension of a lot of chemical processes in the solid state (not only for industry but also for basic science).
1.5 - R EFERENCES
[1] H. Shmalzried, Solid State Reactions, Verlag Chemie, Wheinheim/Bregstrs. 1974.
[2] M. Soustelle, Cinétique hétérogène 3, Ed. Lavoisier - hermès science, Paris, 2006, p. 61- 115 ISBn2-7462-1391-5.
[3] FORCERAM editions SEPTIMA, Paris, 1994. Part : J.C NIEPCE et G. THOMAS, La synthèse par voie solide, chaptres III-2, pages III—9 à III-68.
[4] M. Cournil and G. Thomas (1979) J. Thermal Analysis 17, 435.
[5] G. Thomas and F. Ropital (1985) J. Thermal Analysis 30, 128.
[6] H. Hauffe, Reaktionen in und an fasten Stoffen, 2nd ed., Springer-Verlag, Berlin 1996.
[7] H. Rickert and C. Wagner (1962) Ber. Bunsenges. Phys. Chem. 66, 502.
[8] G. Valensi (1936) C.R. Acad. Sci. Paris 202,309.
[9] R .E. Carter (1961) J. Phys. Chem. 34, 2010.
[10] G. Thomas and F. Ropital (1983) J. Thermal Analysis 28, 91.
[11] F. M. d’Heurle (1988) J. Mater. Res. 3, 167.
[12] R. Bormann (1994) Mater. Res. Soc. Symp. Proc. 343, 169.
[13] K. R. Coffey, Ph.D. thesis, Massachusetts Institute of Technology, Cambridge, 1989.
[14] D. Ingrain et G. Thomas (1982) J. Chim. Phys. 79, 729.
[15] D. Ingrain et G. Thomas (1983) J. Chim. Phys. 80, 199.
[16] U. Gosele and K. N. Tu (1982) J. Appl. Phys. 53, 3252.
[17] U. Gosele and K. N. Tu (1989) J. Appl. Phys. 66, 2619.
[18] G. Thomas and F. Ropital (1985) L'Industrie Céramique 794/5, 335.
[19] F. Ropital and G. Thomas (1985) Elsevier Science Publishers, Amsterdam, 613.
2. T HE H AUNT F OR T HE R EACTION
“Whatever you do will be insignificant, But is very important that you do it .”
(M. Gandhi)
2.1 - S TEP O NE: THE R EACTION
Among solid state reactions, the formation of oxide spinels AB2O4 is at present the most thoroughly investigated [1-3]: first of all because of the relatively simply crystallographic structure of the spinel. Essentially, it consists of nearly close-packed face-centered-cubic subframe of oxygen ions. There are tetrahedral and octahedral interstices of this subframe that are filled by bivalent and trivalent cations, respectively.
In addition, spinels are technically very interesting substances, and one would like to be able to find optimal methods for their preparation. Therefore, the formation of spinels will now be discussed briefly as a classical solid state reaction.
We must first make few assumptions, and under them, it is possible to calculate the rate reaction providing the standard free energy and the component diffusion coefficients of the ions taking part in the reaction: (i) the densities of the reactant oxides AO and B2O3 and the reaction product are equal to their theoretical densities and the contact between them is ideal; (ii) in the quasi-binary system AO-B2O3 only the spinel AB2O4, occurs; (iii) the reaction takes place isothermally; (iv) local thermodynamic equilibrium is maintained within the reaction layer and also at the phase boundaries.
Considering that the transport coefficients of the individual ions are generally rather different from one another and that in spinels the diffusion of oxygen is
negligible, we can eliminate a number of the possible mechanisms. Furthermore, if ideal contact is maintained at the phase boundaries so that the gas phase cannot enter, than the only remaining probable reaction mechanism is the counter-diffusion of cations.
The spinel formation reaction AO + B2O3 AB2O4 is an heterogeneous reaction. Therefore, material must be transported across the phase boundaries and through the reaction product; the resistance to diffusion increases as the thickness of the reaction layer increases. Finally, if the phase boundary resistance can be neglected relative to the diffusional resistance, than local thermodynamic equilibrium will be attained at the phase boundaries. For spinels formation, this state is reached when the reaction layer has attained thickness x of the order of 1 m. Everything we said till now is the theory undergoing the traditional solid state processes for long time of reaction. From this point and for the next chapters in this thesis almost everything is a new and totally unexplored world, from the theory to the experiments used to study this intriguing field that represents the early stages of reaction in the solid state.
2.2 - S TEP T WO: THE O XIDE S PINEL
Minerals of the spine1 group, ideally AB2O4 oxides of diverse chemistry, are widespread in nature and occur in a number of geological environments including, for example, lunar rocks, and meteorites. In the structure of normal spinel, as mentioned before, the A-site is tetrahedrally coordinated and generally occupied by divalent cations (e.g. Mg, Mn, Fe, Ni, Zn). The B-site is octahedrally coordinated and occupied by trivalent cations (e.g. Al, Cr, Fe) or in part by tetravalent, such as Ti. Inverse spine1 is characterized by the occupation of one of the B-sites by the divalent cation with one trivalent cation taking its place on the A-site.
A list of some principle oxide spine1s is given in Table 1:
End-Member Name Abbr Ordering MgAl2O4 Spinel Sp Normal MnAl2O4 Galaxite --- Normal FeAl2O4 Hercynite Hc Normal ZnAl2O4 Gahnite Gn Normal NiAl2O4 --- --- Inverse Table 1: Some members of the principle oxide spinels group
As we are looking for a model to study the first stages of the solid state reaction that leads to the formation of spinel it is necessary to take into account some important remarks: (i) it is necessary to have a well known reaction; (ii) the reaction has to lead to a normal spinel, in this case we can neglect the problem of inverse structure and the percentage of the inversion,; (iii) one of the reactant has to permit to be deposited by means of RF-magnetron sputter and the other one has to be purchased in single crystal with a well known orientation (as we will see later); (iv) the reaction conditions have to be mild. Under these assumption, among the oxide spinels reported in Tab. 1, we selected as potential model reaction the one between Al2O3 single crystal and MO where M = Mn, Fe and Zn.
Let’s consider now the formation of the Mn-Al spinel. First of all we observed, after different deposition attempts, that the stabilization of Mn(II) has been very difficult during the deposition with RF-magnetron sputter and this obliged us to skip to the second candidate.
If we observe the Fe-Al-O system we see a single aluminate based on FeAl2O4
Fig. (2-2): FeO-Al2O3 phase diagram
Fisher and Hoffmann [4] showed that the only stable intermediate oxide in the pseudobinary system FeO-Al2O3 is FeAl2O4 (hecynite). The spinel occurs as a continuous series of solid solutions between FeAl2O4 and Fe3O4. Over a narrower range of composition near FeAl2O4, the spinel equilibrates with “essentially pure Fe” [5,6].
More recently, Meyers et al. [7] determined the composition limits of the FeAl2O4- Fe2O3 spinel solid solution equilibrated with -Al2O3 and with FeO (wüstite) as a function of oxygen partial pressure. At lower oxygen partial pressure their results indicate that -Al2O3 equilibrates with the stoichiometric spinel (FeAl2O4) and essentially pure -Fe equilibrates with spinel of increasing Fe:Al ratio with increasing oxygen partial pressure. Below a critical oxygen partial pressure, the spinel is no longer stable and -Al2O3 equilibrates with -Fe. The corresponding activity of oxygen in -Fe is the critical parameter in predicting spinel formation at internal Fe/ -Al2O3 interfaces.
The minimum oxygen concentration in -Fe necessary for spinel formation at - Fe/ -Al2O3 interfaces increases from approximately 1 at. ppm at 1224 K to 20 at. ppm
approximately 1/10 to 1/5 of the solubility limit of oxygen in -Fe in equilibrium of wüstite. What we can observe from all these data and information is that the conditions to stabilized Fe-Al spinel are very difficult to achieve.
Among the three selected oxide spinels (i.e. ZnAl2O4, MnAl2O4 and FeAl2O4), in conclusion, we selected the couple of parent oxides that leads to a single product, that is Al2O3-ZnO, according to the phase diagram.
Fig. (2.3) – Phase diagram for the couple of parent oxides -Al2O3\ZnO
We choose this couple of parent oxides taking into account that the presence of just one reaction product is a crucial point for the model, in this way we can neglect the variable dues to the presence of different products at different reaction extent and be sure that the only one is the Zn-Al spinel. In addition, ZnO is a very important material in electronic and spintronic fields. A quite large number of topotactical relationships can be found at the interface between ZnO and Al2O3 [8 and references there in] single crystals.
Zinc oxide (ZnO) films are a wide band gap II–VI semiconductors, so they are very useful in light emitting diodes and surface acoustic wave devices. Differences of electron mobility and in the topotactical relationships in ZnO films grown on the
various orientations of sapphire single crystals, have been reported, but there are no systematic studies on the ZnO || Al2O3 interfaces.
To perform the reaction we used Al2O3 single crystals with three different orientations, with the aim to study the mechanism and the kinetic of the process also as a function of the topochemical relationships; the other oxide was powder ZnO that we deposited as a very thin films onto the alumina crystals, as we’ll see more in detail in the next section.
It’s obvious that, once the model and the analysis strategy will be build up, than it will be possible to apply the criterion, with the necessary little modifications, at a lot of solid state reactions and not only to these that concern spinels formation.
2.3 - T HE I DEA
As we said in previous chapter, when we speak about the early stages of solid state reactions we consider that the local chemical equilibrium is not yet attained at the interfaces and that, in these conditions, we have not an adequate probe and model.
It has been demonstrated by our group [9, 10] that insight about mechanism and kinetic at these very early stages can be obtained by performing the reaction using at least one of the two reactants in form of very thin film (10 ÷ 20 nm) and using a local probe for the local chemical environment of one of the constituent.
The approach towards the reactivity in the solid state is primarily based on
relationship between the two reactants, which in turn control formation of the new phase and its spatial and orientational relationships with respect the parent phases. Using one of the reactants in form of film, and considering films of different thicknesses, the ratio between bulk and interfacial free energy can be changed, and the effect of interfacial free energy can be maximized (by using the reactant in form of thin film) and finely tuned. So, it is possible to switch from a situation where the interfacial contribution is the largest to another where bulk contribution rules the process, that is from a quasi–2D to a 3D reactive system. Three different regimes can be explored and clarified.
a) For very thin films, the reactive system is quasi-2D, the interfacial term dominates the overall free energy, and different products can be thermodynamically stabilised by the interfacial free energy: for given structures and compositions, the effect is controlled by the topotactical relationships and is different for different interfaces.
b) For intermediate thickness, once the final phase has nucleated, interfacial movement is the rate-determining step: topotaxy and interfacial free energy control the rate of the reaction.
c) For thick reaction layers, diffusion is the rate-determining step and the interfacial term is negligible.
We expect that the transitions between the regimes could be directed towards different products and investigated by changing the dimensionality of the reactive couple. To the best of our knowledge, this approach towards the reactivity in the solid state is completely original.
The strategy that it has been then performed consists in the use of a lot of different techniques, most of them depending on synchrotron radiation. The principal stages of the strategy of investigation are:
1) The construction of the model, by means of deposition of very thin films of MO (M = divalent metal) on a substrate, that is a single crystal
2) The firing at different temperature and for different amount of time, to carry out the reaction and to have the samples at different extent of reaction
3) The use of M-k edge (where M is the divalent metal) fluo-XAS (but it’s possible to use also refl-EXAFS, depending on the sample) to observe the local environment of the metal
4) The use of M-k edge micro-XANES to have a sort of map of concentration of one of the constituent (that give data that are comparable to those of SIMS)
5) The use of surface and classical diffractions
6) The use of AFM to explore the surface changes that may occur during the reaction
In next chapters and sections we’ll analyze in detail every single point of the strategy we have planned now.
2.4 - R EFERENCES
[1] H. Hauffe, Reaktionen in und an fasten Stoffen, 2nd ed., Springer-Verlag, Berlin 1996.
[2] H. Shmalzried (1965) Ber. Dtsch. Keram. Ges 42, 11.
[4] W. A. Fisher and A. Hoffmann (1956) Arch. Eisenhüttenwes, 27, 343.
[5] L. M. Atlas and W. K. Sumida (1958) J. Am. Chem. Soc. 41, 150.
[6] N. G. Schmahl and H. Dillenburg (1969) Z. Physik. Chem. N. F. 65, 119.
[7] C. E. Meyers, T. O. Mason, W. T. Petuskey, J. W. Halloran and H. K. Bowen (1980) J. Am.
Chem. Soc. 63, 659.
[8] S.-H. Lim, D. Shindo, H-. B Kang and K, Nakamura (2001) J. Cryst. Growth 225, 202.
[9] F. D’Acapito, P. Ghigna, I. Alessandri, A. Cardelli, I. Davoli (2003) Nucl. Instrum. Methods Phys. Res. B 200, 421.
[10] P. Ghigna, G. Spinolo, I. Alessandri, I. Davoli, F. D’Acapito (2003) Phys. Chem. Chem.
Phys. 5 2244.
3. A M ODEL TO E XPLORE E XPERIMENTS AND
T HEORY
“Whenever a theory appears to you as the only possible one, take this as a sign that you have neither understood the theory nor the problem which it was intended to solve .”
(K. Popper)
3.1 - B UILDING UP THE F IRST S TAGES OF THE M ODEL
As we said in previous chapters, when we speak about the early stages of solid state reactions we consider that the local chemical equilibrium is not yet attained at the interfaces and that, in these conditions, we have not an adequate probe and model.
It has been demonstrated by our group [1, 2] that insight about mechanism and kinetics at these very early stages can be obtained by: (i) performing the reaction using at least one of the two reactants in form of very thin film (10 ÷ 20 nm); (ii) using a local probe for the local chemical environment of one of the constituent.
In this section we will explore the experimental procedure that we use to prepare the model to simulate the early stages of the solid state reactions.
After the choice of the two parent oxides, as described in the previous chapter and taking into account all the previous assumptions, we will describe the experimental procedure that it has been carried out for the model preparation. Zinc oxide films were deposited by a RF-magnetron sputtering using a zinc oxide target. The deposition conditions were optimized to give good quality films suitable for the study of the
3.2 - Z nO T HIN F ILMS D EPOSITION
Zinc oxide (ZnO) films are a wide band gap (3.37 eV) II–VI semiconductor, which has a large electro-mechanical coupling factor, high piezoelectric coefficient, photoconductivity, and is transparent in the visible and infrared ranges, so they are very useful in light emitting diodes and surface acoustic wave devices [3]. However, the optical and electrical properties of epitaxial films [4,5] are critically dependent on the defects and interface structure. Srikant et al. [6] reported on the differences of electron mobility in ZnO films grown on the various orientations of sapphire. These properties are attributed to the differences of dislocation density in the films produced as a result of the different lattice mismatches, as a consequence of different orientations of the sapphire. Therefore, it is very important to clarify the interface structure in order to have a full understanding of the epitaxial growth and properties of ZnO films.
Since the lattice mismatch between the ZnO films and the sapphire is low, it is expected that high quality ZnO films on the sapphire will be achieved.
ZnO films can be deposited by a variety of deposition techniques based on the specific applications one needs from these films. Epitaxial ZnO films were grown on sapphire substrates by chemical vapor deposition (CVD) [7]. Recently pulse laser deposition has been tried to deposit ZnO films. By the way, the most common technique to deposit ZnO films, and the one we used, is by sputtering technique. Both DC and RF.
sputtering techniques can be used to deposit ZnO films [8-11]. The resistivity of the deposited ZnO films depends on the sputtering conditions. Usually, the films deposited without oxygen in the sputtering ambient are non-stoichiometric with conducting nature. This is due to the overstoichiometry of Zn in the deposited films [12].
In the present work, as we said, ZnO films were deposited by an RF-magnetron sputtering technique using ZnO target. The deposition conditions were optimized according to Sundaram et al., except for the deposition temperature, because after our preliminary tests it has been observed that on very thin films it’s better to conduct the deposition at room temperature [13]. In fact, as we’ll discuss later, using the deposition conditions reported by Sundaram, we observed that for very thin films at 250 °C we have an initial reaction at the surface. In Fig. (3-1) the XRPD spectra of ZnO reference and ZnO thin film deposited at 250 °C onto Al2O3
( )
1120 as a preliminary test, are reported.2 θ (°)
10 20 30 40 50 60 70
In te ns ity (a rb . u n. )
0 20 40 60 80 100 120
Fig. (3-1): XRPD spectrum of the ZnO thin film deposited at 250 °C onto Al2O3
( )
1120 , according to Sundaram et al. (black line). For comparison purposes the ZnO reference lines are reported (red bars)The ZnO films were deposited by means of a RF-magnetron sputter using a 5 cm diameter by 0.625 cm thick ZnO target (99.999% purity, Sigma-Aldrich). Alumina single square crystals with different orientations, (0001),
(
1102)
and( )
1120 , one side polished (MaTeck) were used as substrates. The substrates were thoroughly cleaned with organic solvent, dried before loading in the sputtering system and placed in thethermocouple attached near the substrate. The system was evacuated to a vacuum of the order of 10-5 bar using in a first step a mechanical pump and then a turbo pump. The deposition has been carried out at ambient temperature.
The argon/oxygen gas injection was achieved through a shower head-like stainless steel tubing with a number of small opening. The ratio of argon (99.99%
purity) and oxygen (99.99% purity) were dialled in the electronic mass flow meters. The gases were mixed before entering the chamber and were allowed to flow through an inlet valve to the chamber. Once the desired vacuum was achieved, a throttle valve was used to adjust the chamber pressure during sputtering. Once the plasma was struck, the pressure and the power were adjusted to the desired conditions. The reflected power was tuned to a minimum value and some times it was necessary to increase the forward power to achieve minimum condition. The substrate was then rotated to place above the target and the deposition was initiated. Table 2 reported a summary of the deposition conditions:
Substrate temperature (°C) 25
Target – substrate distance (cm) 3 Oxygen / Argon flow ratio 2 : 3 Pressure during sputtering (bar) 1 x 10-5
Forward power (W) 60
Reflected power (W) 0
Deposition time (minutes) 10
Thickness (nm) 13
Table 2: Conditions for the sputtering process for samples preparation
